Effects of contralateral noise on envelope-following responses, auditory-nerve compound action potentials, and otoacoustic emissions measured simultaneously

Abstract

This study assessed whether the effects of contralateral acoustic stimulation (CAS) are consistent with eliciting the medial olivocochlear (MOC) reflex for measurements sensitive to outer hair cell (otoacoustic emissions, OAEs), auditory-nerve (AN; compound action potential, CAP), and brainstem/cortical (envelope-following response, EFR) function. The effects of CAS were evaluated for simultaneous measurement of OAEs, CAPs, and EFRs in participants with normal hearing. Clicks were presented at 40 or 98 Hz in three ipsilateral noise conditions (no noise, 45 dB SPL, and 55 dB SPL). For the no noise condition, CAS suppressed or enhanced EFR amplitudes for 40- and 98-Hz clicks, respectively, while CAS had no significant effect on CAP amplitudes. A follow-up experiment using slower rates (4.4–22.2 Hz) assessed whether this insignificant CAS effect on CAPs was from ipsilateral MOC stimulation or AN adaptation; however, CAS effects remained insignificant despite favorable signal-to-noise ratios. CAS-related enhancements of EFR and CAP amplitudes in ipsilateral noise were not observed, contrary to the anti-masking effect of the MOC reflex. EFR and OAE suppression from CAS were not significantly correlated. Thus, the effects of CAS on EFRs may not be solely mediated by the MOC reflex and may be partially mediated by higher auditory centers.

I. INTRODUCTION

The auditory system includes afferent and efferent neural pathways which are responsible for carrying auditory information between the ears and brain. The outer hair cells (OHCs) receive efferent feedback from the brain through the medial olivocochlear (MOC) bundle (Rasmussen, 1946). The MOC reflex inhibits cochlear amplifier gain by reducing OHC motility, and it is thought to play a role in protection from loud noise exposure as well as facilitate speech understanding in background noise (Guinan, 1996). In a quiet background, eliciting the MOC reflex decreases the firing rate of single auditory-nerve (AN) fibers to sound (e.g., Wiederhold, 1970). However, for transient sounds presented in background noise, eliciting the MOC reflex improves the signal-to-noise ratio (SNR) between a target sound and background noise, producing an anti-masking effect (e.g., Nieder and Nieder, 1970).

In humans, the MOC reflex is often studied using contralateral suppression of otoacoustic emissions (OAEs), which are sounds present in the ear canal as a by-product of OHC motility and provide insight on the health of the cochlear amplifier (Berlin et al., 1993). Auditory evoked electrical potentials complement OAEs by extending the assessment of the MOC reflex to neural responses of the AN, auditory brainstem, and cortex. An understanding of the effects of the MOC reflex on the neural response of the auditory system may provide insight on how the MOC reflex influences auditory perception (Smith et al., 2017a; Smith et al., 2017b).

The presentation of contralateral noise often suppresses the response of OAEs (e.g., Hood et al., 1996) as well as the envelope-following response (EFR; e.g., Galambos and Makeig, 1992), which is a brainstem/cortical auditory evoked potential that phase-locks to the acoustic temporal envelope (Purcell et al., 2004). This suppression is consistent with contralateral acoustic stimulation (CAS) evoking the MOC reflex. A previous study (Mertes and Leek, 2016) compared an EFR-based measure of MOC activity with a measure based on transient-evoked otoacoustic emissions (TEOAEs). Mertes and Leek (2016) proposed that the EFR may be a promising tool for studying MOC activity, particularly in listeners with cochlear hearing loss, because contralateral suppression of AN responses has been reported to be larger than contralateral suppression of OAEs in animal and human studies (Puria et al., 1996; Lichtenhan et al., 2016). Mertes and Leek (2016) simultaneously measured TEOAEs and EFRs evoked by a 40-Hz click train and predicted that the introduction of contralateral broadband noise (CBBN) would decrease the amplitude of both measures, consistent with eliciting the MOC reflex and causing an inhibition of cochlear amplifier gain. Their results showed a significant suppression of EFR and TEOAE amplitudes in response to CBBN; however, the magnitude of EFR and TEOAE suppression was not significantly correlated, which is inconsistent with the suppression of the EFR originating entirely from stimulation of the MOC reflex.

Studies in laboratory animals suggest that the primary generators of the EFR for low modulation frequencies (e.g., <80 Hz) are located within the auditory brainstem and cortex (Kuwada et al., 2002). It is possible that processes influencing these generators contribute to the contralateral suppression of the EFR. Simultaneous measurement of the effects of CBBN on the EFR and an earlier neural measurement, the AN compound action potential (CAP)—which does not originate in the brainstem and cortex—may reveal the extent to which contralateral suppression of the EFR is consistent with eliciting the MOC reflex.

The CAP results from the summed activity of several AN fibers firing synchronously and can be elicited by a transient stimulus such as a click, chirp, or brief tone burst. Recently, Chen and Jennings (2022) showed that the envelope of a stimulus can evoke a CAP in response to each cycle of modulation; they called this potential “CAP_ENV.” The differences between CAP_ENV and the EFR are the putative neural generators (i.e., AN vs brainstem/cortex) and the location of the active electrode (i.e., eardrum vs forehead). For this article, CAP_ENV refers to the waveform produced by a click train—as observed from an eardrum electrode—which results in a CAP to each click in the train. In this study, we simultaneously evoked CAPs_ENV, EFRs, and TEOAEs in response to a train of clicks presented in a quiet background to test the hypothesis that the presentation of CBBN suppresses CAP_ENV and EFR amplitudes similarly, which is consistent with reduced AN and brainstem/cortical responses from eliciting the MOC reflex. Furthermore, we hypothesize that click-evoked CAP_ENV and EFR amplitudes measured in the presence of ipsilateral noise will be enhanced with the introduction of CBBN, consistent with the anti-masking effects of the MOC reflex. Finally, we compare the effects of CBBN on CAP_ENV and EFR amplitudes with the effects on TEOAE amplitudes.

II. METHODS

A. Participants

The participants in this study included 12 young-adults (8 female and 4 male) with normal hearing and ages ranged from 19 to 26 years old (mean = 23.25 years of age). Inclusion criteria involved normal otoscopic evaluation, normal middle ear function (type A tympanograms), and normal (≤20 dB hearing level; HL) hearing thresholds from 250 to 8000 Hz. Normal middle ear function was defined as ear canal volume between 0.9 and 2.5 cm³, compliance between 0.3 and 1.7 mmhos, and peak pressure of ±100 daPa (Margolis and Hunter, 1999). Participants were recruited from the student body at the University of Utah. The procedures of this study were approved by the University of Utah Institutional Review Board, and written informed consent was obtained from all participants prior to their enrollment in the study. Eleven of these participants were compensated for their time at a rate of $15/h, and one participant was a volunteer.

B. Stimuli

The stimuli of this study were designed to closely follow those used in Mertes and Leek (2016). The sampling rate of the stimulus generation system [RZ6, Tucker-Davis Technologies (TDT), Alachua, FL] was 24 414.0625 Hz. The stimulus was a train of 80-μs clicks presented at 75 dB peSPL (peak equivalent sound pressure level) at two stimulus repetition rates. Mertes and Leek (2016) used a stimulus rate of 39.0625 Hz (∼40 Hz), which was selected to yield an integer number of samples for their 24 414.0625 sampling rate. We chose the same repetition rate (39.0625 Hz; ∼40 Hz) and an additional rate of 97.65625 (∼98 Hz), which also yields an integer number of samples.

The click train stimuli were presented to the right ear while CAS, in the form of CBBN, was presented to the left ear at 60 dB A (∼61.5 dB SPL), which is consistent with that in Mertes and Leek (2016). This CAS intensity level is an effective elicitor of the MOC reflex and limits the contribution of the middle ear muscle reflex (MEMR; Guinan et al., 2003; Mertes and Goodman, 2016). A schematic of an interleaved stimulus recording is displayed in Fig. 1.

FIG. 1.

(Color online) Schematic of one interleaved presentation of the stimuli. Click trains were presented to the right ear in alternating polarity and are displayed in the top half of the panel (red waveform). The number of clicks is reduced by a factor of 10 to aid visualization of the click train. The gray waveform in the top panel (right ear) represents ipsilateral broadband noise (when present). CBBN was presented to the left ear and is shown in the bottom half of the panel (blue waveform). Two stimulus conditions (−CAS and +CAS) are depicted and separated by a 2-s interval of CBBN presented alone.

In addition to recording measurements at two different stimulus repetition rates, this study included measurements in the presence and absence of ipsilateral background noise (IBBN). The following experimental conditions were measured for both repetition rates (40 and 98 Hz): no ipsilateral noise, 45 dB SPL IBBN, and 55 dB SPL IBBN. The calibrations of the click train, IBBN, and CBBN were achieved using a 2 cc coupler (AEC203, PCB Piezotronics, Depew, NY) and a sound level meter (System 824, Larsen-Davis, Provo, UT).

C. Procedures and equipment

Data for all experimental conditions were collected in a sound-treated booth in the Auditory Perception and Physiology Research Laboratory at the University of Utah. Before and after each session, otoscopic examination and tympanometry were performed. Immittance measures were obtained using a Grayson-Stadler Industries (GSI, Eden Prairie, MN) 39 tympanometer. For the initial hearing test, pure-tone thresholds were measured using a GSI 61 clinical audiometer.

Two-channel auditory evoked potential recordings were obtained to compare AN potentials with brainstem/cortical potentials. TEOAEs were recorded simultaneously with evoked potentials using an ER7C (Etymotic Research, Elk Grove, IL) probe microphone, which was electromagnetically shielded (Simpson et al., 2020). Electrophysiology recordings were obtained using TDT hardware and software (Alachua, FL). The hardware included a high-performance workstation (TDT-WS4), multi-input/output signal processor (TDT-RZ6), low-impedance head stage (TDT-RA4LI), bio-amplifier (TDT-RA4PA), and electromagnetically shielded insert earphones (ER-3C, Etymotic Research, Elk Grove, IL). TDT Synapse software was used to record evoked potentials and controlled via custom matlab (The Mathworks, Natick, MA) code and the Synapse Application Program Interface.

D. Participant preparation

To ensure low impedances at the electrode sites, the participant's skin was cleaned with an alcohol wipe and lightly abraded with gauze and NuPrep gel (Bio-Medical Instruments, Clinton Township, MI). AgCl disk surface electrodes were filled with Ten20 conductive paste (Bio-Medical Instruments, Clinton Township, MI) and secured to the participant's scalp using medical tape. An active electrode was placed on the high forehead at the intersection of the midline and skin just below the hairline to measure the EFR. The ground and inverting (reference) electrodes were placed on the nasion and ipsilateral earlobe, respectively. A second active electrode was placed on the tympanic membrane (TM) of the right ear to measure CAP_ENV, which is the steady-state version of the CAP and has a similar relationship as the EFR does to the auditory brainstem response (ABR). The TM electrode was custom built and is associated with greater participant comfort and greater CAP amplitudes compared to a commercially available electrode (Simpson et al., 2020). Prior to placing the TM electrode, the right external ear canal was bathed with warm saline for 1 min and then drained. Next, the TM electrode was placed on the right TM under visualization with an otoscope and secured in place with medical tape and a foam insert. A shielded jumper cable (Intelligent Hearing Systems, Miami, FL) was used to connect the TM electrode to the bio-amplifier to limit electrical artifact. A schematic of the equipment setup is shown in Fig. 2.

FIG. 2.

(Color online) Schematic of the equipment setup for simultaneous measurements of TEOAEs, CAPs_ENV, and EFRs. Transducers (ER3Cs) for the probe (click train) and CAS (broadband noise) are displayed near the top of the right panel. Reference (Ref.) and ground (Gnd.) electrodes are denoted by red and green lines while active electrodes are shown by purple [TM, channel 1 (Ch. 1)] and yellow [high forehead, EFR, channel 2 (Ch. 2)] lines, respectively. The panel to the left depicts a zoomed view of the setup for the right ear, which includes a microphone (Mic.) for measuring ear canal sound pressure.

The 80-μs clicks were presented at 75 dB peSPL to the participants right ear for conditions with and without CAS. The 60 dB A CBBN was presented to the left ear in conditions with CAS. The condition with no CAS was presented for 32 s with the polarity of the click train alternating every 2 s. This was followed by a 2-s period of only CBBN to allow for the full onset of the MOC reflex (Backus and Guinan, 2006; Mertes and Leek, 2016). Then, the measurement was repeated in the presence of CAS for 32 s, followed by 2 s of silence at the end to allow for a full offset of the MOC reflex (Backus and Guinan, 2006; Mertes and Leek, 2016). The with/without CAS conditions were interleaved for ten blocks for measurements with both click rates (40 and 98 Hz) and all ipsilateral noise conditions (no noise, 45 dB SPL, and 55 dB SPL). This stimulus presentation paradigm was designed to closely follow that of Mertes and Leek (2016) except for the use of alternating polarity of the probe stimulus. Alternating polarity was used to limit the contribution of the cochlear microphonic (CM) to CAP_ENV and EFR recordings.

To maintain a consistent state of arousal throughout the recordings, the participants wore a vibrating motor on their ankle, which turned on every 90 s (Jennings and Aviles, 2023). The participant could then turn the motor off by pressing a button. If the vibration was not turned off by the participant within 30 s, the experimenter would pause the recordings and enter the sound booth to ensure that the participant was awake. The vibrating motor was used throughout all recordings.

E. Data preprocessing

The following process, similar to that of Mertes and Leek (2016), was applied to CAP_ENV and EFR recordings to determine the effects of CAS on these evoked potentials. Each 68-s sweep was divided into thirty-four 2-s epochs and sorted into −CAS (16 epochs), +CAS (16 epochs), silence (1 epoch), and noise-only (1 epoch) matrices. The −CAS and +CAS matrices were further divided into sub-matrices for condensation and rarefaction clicks. These matrices were combined for all 10 sweeps; thus, the −CAS and +CAS matrices consisted of 160 epochs (80 condensation and 80 rarefaction), and the silence and noise-only matrices consisted of 10 epochs each. Each 2-s epoch was bandpass filtered with a 1024-point Hann window after applying 50-ms cos² onset/offset ramps. Filter cutoff frequencies were 30 and 50 Hz for recordings associated with the 40 Hz click rate, whereas they were 73 and 123 Hz for recordings associated with the 98 Hz click rate. To improve frequency resolution, −CAS and +CAS matrices were reshaped into two 16-s buffers, each associated with a different polarity (i.e., condensation or rarefaction). The signal was defined as the average of the 16-s buffers across the ten sweeps. To compute the noise floor, every other sweep was inverted, and all sweeps were averaged. The signal's fast Fourier transform (FFT) was computed over the duration of the signal (16 s, rectangular window) and scaled to microvolts for the −CAS and +CAS conditions in condensation and rarefaction polarities. This process was repeated for the average silence and noise-only buffers. The signal's magnitudes for the −CAS and +CAS conditions were defined as the maximums of the absolute values of the averages of the FFTs for the condensation and rarefaction polarities, where the maximum was computed within a narrow frequency window that included the click rate (37–41 Hz for 40 Hz click rate and 95.7–99.7 Hz for 98 Hz click rate). This process was repeated to obtain the magnitude of the noise floor, except the noise magnitude was defined as the average rather than the maximum value within the narrow frequency window. The change in CAP_ENV (or EFR) amplitude resulting from the presentation of CAS was defined as the CAP_ENV (or EFR) amplitude in the −CAS condition minus the CAP_ENV (or EFR) amplitude in the +CAS condition. Although phase changes may have occurred from presenting CAS, such changes were not considered in this report as the hypotheses centered on evaluating changes in magnitude.

The following process, similar to that in Mertes and Leek (2016), was applied to determine the effects of CAS on TEOAEs. Each 68-s sweep was divided into thirty-four 2-s epochs, as was performed for the CAP_ENV and EFR recordings. The sample of the first negative peak of the click stimulus for a given 2-s epoch was determined from known delays associated with the electronic stimulus, equipment processing, and the tube of the ER-3C earphones. These delays totaled 5.7 ms, which corresponded to sample 140. This sample defined the start of the first click in each 2-s epoch. Samples defining the start of the remaining clicks were identified from the period of the click rate (e.g., 40 Hz). A 20-ms analysis window (10 ms for the 98-Hz condition) was drawn from the start of each recorded click. The first 3.5 ms of this window was set to zero to remove the stimulus and retain the TEOAE waveform for each presented click. These TEOAE waveforms were ramped with 2.5 ms onset/offset ramps and bandpass filtered from 1000 to 4000 Hz (FIR filter and 256-point Hann window). Waveforms whose root mean square (rms) amplitudes were two standard deviations above the mean rms amplitude were discarded. Artifact-free TEOAE waveforms were sorted into +CAS and −CAS conditions, and separate averages for condensation and rarefaction clicks were computed for each condition. The signal was defined as the mean difference between the average condensation and rarefaction TEOAE waveforms, whereas the noise floor was defined as the mean of the average condensation and rarefaction waveforms. A 2048-point FFT was computed and expressed as a power spectrum in dB SPL for the −CAS and +CAS conditions for the signal and noise floor waveforms, and the SNR within each frequency bin was computed. Suppression of TEOAEs was computed for each frequency bin as the difference in signal power between conditions (i.e., +CAS and −CAS). Total TEOAE suppression was defined as the sum of TEOAE suppression across all frequency bands with SNR > 6 dB. This total TEOAE suppression value was expressed in dB for each participant. TEOAE suppression was computed only for measurements that did not include ipsilateral noise, as measurements with IBBN did not meet the SNR criterion.

F. Statistical analysis

A repeated-measures analysis of variance (ANOVA) with three repeated measures (CAS, +CAS, −CAS; ipsilateral noise, none, 45, and 55 dB SPL; repetition rate, 40 and 98 Hz) was conducted for the CAP_ENV and EFR data separately. A similar ANOVA was conducted for the TEOAE data (CAS, +CAS and −CAS; repetition rate, 40 and 98 Hz). Post hoc testing included paired t-tests that were adjusted for multiple comparisons using Bonferroni's method. Significance was evaluated at α = 0.05.

G. Follow-up experiment using lower click rates

The relatively high click rates of 40 and 98 Hz can result in stimulation of the ipsilateral MOC reflex, which may limit any additional MOC effects elicited by CAS. For example, Boothalingam and Purcell (2015) showed that TEOAE magnitudes are reduced for clicks with rates ≥ 31.25 Hz compared to a baseline condition where clicks were presented at approximately 1.25 Hz, which is consistent with eliciting the ipsilateral MOC reflex. Furthermore, changes in CAP and CM magnitudes from electrical stimulation of the olivocochlear bundle depend on ipsilateral stimulus presentation rate for experiments in chinchillas (Elgueda et al., 2011). The mechanisms of these stimulus rate effects are unknown but may be related to neural adaptation for CAP measurements. An additional experiment involving 12 participants with normal hearing was completed to evaluate the influence of click rate on the effects of CAS. This experiment involved simultaneously measuring CAPs_ENV and EFRs in response to a 2.25 s train of upward-frequency chirps presented at 60 dB peSPL. The interval between adjacent chirps within the train was 227 ms (4.4/s), 90 ms (11.1/s), or 45 ms (22.2/s). A 60 dB SPL, 750 ms CBBN (i.e., CAS) was gated on 750 ms after the onset of the chirp train. Thus, the first and last thirds of the chirp train did not include presentation of CAS. This stimulus presentation paradigm is similar to that in Jennings and Aviles (2023), who studied the effects of CAS on the CM, and attempts to capture real-time changes in response magnitude as a result of eliciting the MOC reflex (Liberman et al., 1996). Recordings were obtained for 700 sweeps of the 2.25-s chirp train in alternating polarity. Chirp trains were presented continuously [i.e., interstimulus interval (ISI) = 0 ms], resulting in an ISI of 1.5 s for presentation of CAS. The spectrotemporal properties of the chirp were identical to those in Chertoff et al. (2010). Specifically, chirp frequencies spanned 450–10 000 Hz, and the instantaneous frequency of the chirp followed the basilar membrane delay estimates from derived-band CAPs from Eggermont (1979). The effect of CAS was defined as the difference in the average peak-to-peak amplitude for chirp-evoked CAPs_ENV (or EFRs) before CAS onset (i.e., <750 ms) and CAPs_ENV (or EFRs) during CAS (i.e., 750–1500 ms).

III. RESULTS

As displayed in Fig. 3, the presentation of CAS suppressed the magnitude of the EFR for 40-Hz clicks for 11 out of 12 participants, whereas the opposite effect (i.e., magnitude enhancement) was observed for 98 Hz clicks for 10 out of 12 participants. The presentation of CAS did not have a consistent effect on CAP_ENV magnitude for either 40- or 98-Hz clicks (Fig. 4).

FIG. 3.

(Color online) EFR magnitude in response to clicks presented at 40 Hz (left) and 98 Hz (right) for individual subjects (x axis). Each panel shows EFR magnitude in the −CAS (black- or red-filled bars) and +CAS (white-filled bars) conditions. The noise floors for each participant are displayed as gray bars.

FIG. 4.

(Color online) CAP_ENV magnitude in response to clicks presented at 40 Hz (left) and 98 Hz (right) for individual subjects (x axis). Each panel shows CAP_ENV magnitude in the −CAS (black- or red-filled bars) and +CAS (white-filled bars) conditions. The noise floors for each participant are displayed as gray bars.

Figure 5 shows the group-average EFR magnitudes for several IBBN levels for clicks presented at 40- (black-bordered bars) and 98-Hz repetition rates (red-bordered bars) in the −CAS (filled bars) and +CAS (open bars) conditions. All main effects, two-way, and three-way interactions were significant [CAS, F(1,11) = 10.1, p = 0.009; modulation frequency, F(1,11) = 39.9, p < 0.001; ipsilateral noise level, F(2,22) = 45.4, p < 0.001; CAS ∗ modulation frequency, F(1,11) = 34.5, p < 0.001; CAS ∗ ipsilateral noise level, F(2,22) = 6.0, p = 0.009; modulation frequency ∗ ipsilateral noise level, F(2,22) = 13.1, p < 0.001; CAS ∗ modulation frequency ∗ ipsilateral noise level, F(2,22) = 9.2, p = 0.001]. As expected, EFR magnitudes significantly decreased in the presence of 45 dB SPL [t(11) = 3.03, p = 0.011] and 55 dB SPL [t(11) = 11.01, p < 0.001] IBBN compared to quiet. We found that for 40-Hz clicks, EFR magnitudes decreased with the presentation of CAS in all IBBN level conditions. The decrease in 40-Hz EFR magnitude was statistically significant in quiet [t(11) = 4.65, p = 0.001] and 45-dB SPL IBBN [t(11) = 4.33, p = 0.001] but was not statistically significant in the 55-dB SPL IBBN condition [t(11) = 2.31, p > 0.05] after correcting for multiple comparisons. For the 98-Hz clicks, EFR magnitude significantly increased with the presentation of CAS in quiet [t(11) = –4.18, p = 0.002] and 55-dB SPL IBBN [t(11) = 3.07, p = 0.006]. A similar increase was observed for the 45-dB SPL IBBN; however, this increase was not significant after correcting for multiple comparisons.

FIG. 5.

(Color online) Group-average EFR magnitudes as a function of ipsilateral broadband noise level for clicks presented at 40- (black-bordered bars) and 98-Hz (red-bordered bars) repetition rates in the −CAS (filled bars) and +CAS (open bars) conditions. The group-averaged noise floor is depicted by the gray bars. Error bars are one standard error of the mean. Asterisks represent significant comparisons for α < 0.01.

Figure 6 shows the group-averaged CAP_ENV magnitude as a function of IBBN level for clicks presented at 40- (black-bordered bars) and 98-Hz repetition rates (red-bordered bars) in the −CAS (filled bars) and +CAS (open bars) conditions. The main effect of IBBN was significant [F(2,22) = 13.7, p < 0.001] as CAP_ENV magnitudes were significantly lower in the presence of 45 dB SPL [t(11) = 3.43, p = 0.006] and 55 dB SPL [t(11) = 4.15, p = 0.002] IBBN compared to CAP_ENV amplitudes measured in quiet. We found that for the 40 and 98 Hz clicks, CAP_ENV magnitude was not significantly influenced by the presentation of CAS for any IBBN condition.

FIG. 6.

(Color online) Group-average CAP_ENV magnitudes as a function of ipsilateral broadband noise level for clicks presented at 40- (black-bordered bars) and 98-Hz (red-bordered bars) repetition rates in the −CAS (filled bars) and +CAS (open bars) conditions. The group-averaged noise floor is depicted by the gray bars. Error bars are one standard error of the mean.

Consistent with previous literature (e.g., Berlin et al., 1993; Mertes and Leek, 2016), significant contralateral suppression of TEOAEs was observed for the 40- [t(11) = 3.73, p = 0.005] and 98-Hz [t(11) = 4.35, p = 0.002] repetition rates as depicted in Fig. 7. The main effect of CAS on OAE amplitudes was statistically significant (F[1,9] = 17.4, p = 0.002), but this effect did not depend significantly on the click rate, as shown by an insignificant interaction effect (F[1,9] = 3.6, p = 0.09). Suppression of TEOAEs could not be quantified in two subjects in the 40-Hz condition because of poor SNRs. A correlation analysis, displayed in Fig. 8, revealed that EFR and TEOAE suppression (Δ dB) were not significantly related for either repetition rate (40-Hz clicks, r = 0.05, p = 0.54; 98-Hz clicks, r = 0.14, p = 0.23), which is consistent with previous literature (Mertes and Leek, 2016).

FIG. 7.

(Color online) Magnitudes [(A) and (B)] and suppression [(C) and (D)] of TEOAEs for individual participants (x axis) measured using 40- (left panels, black) and 98-Hz click trains (right panels, red). Suppression could not be computed in two subjects in the 40-Hz condition due to poor SNRs.

FIG. 8.

(Color online) Scatterplots and regression lines for EFR and TEOAE suppression for data obtained with the 40- (top panel, black) and 98-Hz (bottom panel, red) clicks.

Suppression of CAP_ENV amplitudes was not observed for measurements made with high (i.e., 40 and 98 Hz) click rates, which are possibly the result of stimulation of the ipsilateral MOC reflex or AN firing-rate adaptation. This possibility was addressed in the follow-up experiment, where slower rates were used. The effects of CAS on CAPs_ENV and EFRs evoked by trains of rising-frequency chirps presented at 4.4, 11.1, and 22.2 Hz are displayed in Fig. 9. The top three panels display group-averaged (N = 12) waveforms for CAPs obtained in response to 4.4/s [Fig. 9(A)], 11.1/s [Fig. 9(B)], and 22.2/s [Fig. 9(C)] chirps. These waveforms are enclosed within a time-varying estimate of the standard error of the mean (cyan shading). The gray area between 750 and 1500 ms marks the interval when CAS was present. Average CAP_ENV amplitudes during baseline and CAS periods were statistically equivalent for all chirp rates (4.4 Hz, t[11] = 0.96, p = 0.36; 11.1 Hz, t[11] = 1.18, p = 0.26; 22.2 Hz, t[11] = 0.86, p = 0.41). In contrast, average EFR amplitudes for the fastest rate (22.2/s) were significantly decreased (t[11] = 3.19, p = 0.009) during CAS compared to baseline [Fig. 9(D)]. Amplitudes for EFRs recorded for the other rates were not significantly reduced by CAS (4.4 Hz, t[11] = 1.69, p = 0.12; 11.1 Hz, t[11] = 0.6, p = 0.56).

FIG. 9.

(Color online) Grand average (N = 12) CAP_ENV [(A), (B), and (C)] and EFR (D) waveforms in response to trains of 60 dB SPL chirps presented at 4.4 (A), 11.1 (B), and 22.2 Hz [(C) and (D)]. The gray shaded rectangle in each panel marks the presentation of CAS. Shaded areas circumscribing the waveforms represent ±1 standard error of the mean. EFR waveforms for the 4.4- and 11.1-Hz chirp trains are not displayed because the effects of CAS were not significant for these frequencies.

IV. DISCUSSION

The amplitude of the 40-Hz EFR measured in a quiet background was suppressed by CAS in most (11 out of 12) participants (Fig. 3); however, no such suppression was observed for the amplitude of the CAP_ENV for the original (40- and 98-Hz probes), or follow-up (4.4-, 11.1-, and 22.2-Hz probes) experiments. The lack of CAP_ENV suppression does not support our hypothesis that CAP_ENV and EFR amplitudes would be similarly suppressed by CAS. Similarly, our hypothesis of CAS-related enhancements of the amplitudes (i.e., anti-masking effect) of the 40-Hz EFR and CAP_ENV measured in ipsilateral noise was not supported by the data. Instead, we found that these amplitudes were suppressed (EFR) or unchanged (CAP_ENV) by CAS when measured during the presentation of 45- or 55-dB SPL ipsilateral noise. Thus, our data are inconsistent with the anti-masking effect of the MOC reflex observed from AN responses of laboratory animals (Nieder and Nieder, 1970; Kawase et al., 1993; Kawase and Liberman, 1993). These important results highlight the difficulty of translating several compelling animal research studies to humans on how the MOC reflex affects AN responses. Although CAS-related enhancements were observed for 98-Hz EFRs, these enhancements occurred with and without the presentation of ipsilateral noise. An enhancement of 98-Hz EFR amplitudes by CAS for measurements in a quiet background are inconsistent with eliciting the MOC reflex as eliciting this reflex is expected to suppress AN activity when measured in a quiet background (Liberman, 1989). Finally, we confirmed a lack of correlation (Mertes and Leek, 2016; Mertes and Potocki, 2022) between the magnitudes of TEOAE and EFR suppression, which does not support the hypothesis that suppression of the 40-Hz EFR by CAS is mediated by the MOC reflex. Conclusions drawn from these findings must be tempered by challenges of obtaining high SNRs, the effects of the stimulus presentation paradigm, and potential challenges in replicating small-sized effects among different laboratories when assessing the effects of CAS on the CAP. Moreover, conclusions must consider the possibility that suppression of the 40-Hz EFR may be mediated by collateral branches of the MOC system within the brainstem, other brainstem processes, and higher auditory centers (e.g., thalamus and auditory cortex) as discussed below.

A. Effects of CAS on CAPs in humans

The lack of an effect of CAS on CAP_ENV is unexpected given that suppression of the CAP by contralateral noise is well-established in animal literature, including studies in cats (Liberman, 1989; Kawase et al., 1993; Kawase and Liberman, 1993; Puria et al., 1996), chinchillas (Elgueda et al., 2011), guinea pigs (Popelar et al., 2001), and gerbils (Huang et al., 1994). These CAS effects are typically measured for transient tonal probes presented at low levels (e.g., <50 dB SPL); however, smaller yet significant suppression has also been reported for clicks presented at moderate levels (i.e., 60–80 dB peSPL; Elgueda et al., 2011).

Study of the effects of CAS on the human CAP has been limited to relatively high-level (e.g., >50–60 dB SPL) clicks (Lichtenhan et al., 2016), chirps (Smith et al., 2017b), filtered clicks (Folsom and Owsley, 1987), and tone bursts (Najem et al., 2016) as CAPs measured at lower levels are absent or have poor SNRs. Importantly, moderate-to-high level (e.g., >60 dB SPL) sounds are associated with upward spread of mechanical and, thus, neural excitation along the cochlear length (Lee et al., 2019; Goodman et al., 2021) and less cochlear amplifier gain compared to lower levels (Cooper, 2004). Thus, suppression of CAPs from presentation of CAS may be small at these levels given that responses may be dominated by passive cochlear processing. Further, CAP suppression by CAS might be weak at these levels given the absence of MOC effects on OAEs measured beyond 3–4 kHz (Goodman et al., 2013) and the contribution of the cochlear base to CAPs (Eggermont, 1976). Finally, CAP suppression might be smaller in humans compared to animals given the relatively sparse MOC innervation density reported for human temporal bones (Liberman and Liberman, 2019).

The effects of CAS on OAEs and CAPs are typically on the order of a few decibels. Accurate detection of these small effects depends strongly on SNR (Goodman et al., 2013). The mean and standard deviation for SNRs measured in the current study for CAP_ENV in the 40-Hz quiet condition (Fig. 4) were 16.6 dB and 3.4 dB, respectively. The average change in CAP_ENV amplitude from the presentation of CAS was +0.40 dB. According to Goodman et al. (2023), an experiment with 12 participants and 1 repeated measurement per participant should be sensitive to an ∼2 dB change in CAP amplitude if the SNR is ∼25 dB. Thus, the SNR may have been too poor to observe a change in CAP_ENV amplitude for data collected with the 40-Hz probe. Conversely, the mean SNR measured in the follow-up experiment (Fig. 9) was 27 dB for the 4.4-Hz chirps. This SNR is larger than the SNRs for Lichtenhan et al. (2016; 26 dB for clicks) and Smith et al. (2017b; 22.2 dB for chirps), suggesting that SNR does not explain the lack of CAP_ENV suppression in the current study for the expected effect size (i.e., 1–2 dB). Of the 12 participants in the follow-up study, 1 exhibited significant suppression of CAP_ENV for all 3 rates, another exhibited significant suppression for 2 rates (11.1 and 22.2 Hz), and 3 other participants exhibited suppression for 1 rate. The average suppression for these cases was 4.9% (–0.42 dB). Thus, only 25% of measurements resulted in significant suppression, and the magnitude of this suppression was less than half a decibel. Given this small effect size and the relative scarcity of participants with significant suppression, it is possible that significant CAP_ENV suppression may have been observed with a larger sample size for the follow-up study.

The original and follow-up experiments measured the effects of CAS for CAPs_ENV measured across a wide range of click rates (4.4–98 Hz). As click rate is increased, the effects of AN adaptation and ipsilateral MOC stimulation are expected to increase and may limit any additional firing-rate suppression from eliciting the contralateral MOC reflex with CAS. Average peak-to-peak CAP_ENV amplitudes during the baseline period decreased from 2.1 to 1.7 μV as the click rate increased from 4.4 to 22.2 Hz (follow-up experiment), which is consistent with AN adaptation. Despite these adaptation effects, suppression of CAP_ENV was not observed during the presentation of CAS for any click rate, suggesting that the effects of click rate are not responsible for the lack of CAP_ENV suppression in the current study. The stimulus presentation paradigm consisted of interleaved blocks of silence and CAS in the contralateral ear. The blocks of silence between presentations of CAS were relatively long (Fig. 1, every 32 s, original experiment) or short (1.5 s, follow-up experiment) as was the duration of the CAS (32 s, original experiment; 750 ms, follow-up experiment). These CAS presentation paradigms could have facilitated a slow buildup of CAP suppression associated with MOC “slow” effects (e.g., Sridhar et al., 1995), and this buildup could have persisted during the interleaved blocks of silence. Sustained MOC activity across blocks of silence and CAS is expected to limit the ability to detect MOC effects when comparing CAP amplitudes across blocks. Support for the potential for MOC slow effects in this study comes from Larsen and Liberman (2009), who showed that slow effects build over 2–3 min before stabilizing, and Liberman (1988), who showed that slow effects may not fully dissipate until 5 min after the offset of a 10-min-long presentation of contralateral noise. Nevertheless, significant suppression of TEOAE amplitudes were observed for 40- and 98-Hz click rates (Fig. 7), which does not support the argument that MOC slow effects are responsible for the lack of CAP_ENV amplitude suppression. The magnitude of CAP suppression by CAS is typically less than 1.5 dB (Lichtenhan et al., 2016; Smith et al., 2017b). Our findings suggest that the small size of this effect may make CAP suppression by CAS difficult to replicate across laboratories that often use different stimulus and recording paradigms. Indeed, Smith et al. (2017b) reported significant suppression of CAPs evoked by chirps but not by clicks measured for the same participants, thereby illustrating the lability of demonstrating significant CAP suppression by CAS.

B. Effects of CAS on EFRs

Previous studies have implied or hypothesized that suppression of the 40-Hz EFR may be mediated by the MOC reflex (Ozdamar and Bohorquez, 2008; Maki et al., 2009; Mertes and Goodman, 2016). This hypothesis is consistent with the effects of the MOC reflex on “central masking” as measured from the audiograms of macaque monkeys for whom the MOC bundle was intact or resected (Smith et al., 2000). Central masking is defined as an increase in detection thresholds in the presence of CAS compared to those without CAS. Smith et al. (2000) found that for monkeys with MOC bundles intact, CAS resulted in an average increase of 4–6 dB for audiometric thresholds for 1000–4000 Hz tones compared to no CAS. For a monkey with a resected MOC bundle, such shifts in threshold were observed before but not after the MOC bundle was resected. Contralateral suppression of the 40-Hz EFR and central masking are consistent with a reduction in cochlear amplifier gain via the MOC reflex to the extent that MOC effects are inherited by central neurons, and these neurons contribute to the 40-Hz EFR and perception (i.e., detection of tones).

Despite the hypothesized role of the MOC reflex in the suppression of the 40-Hz EFR, several findings are inconsistent with a MOC-mediated decrease in the EFR at this frequency. First, a CAS-related decrease in cochlear amplifier gain from eliciting the MOC reflex is expected to reduce AN firing rates as well as the firing rates of subsequent afferent neurons in the brainstem, midbrain, and cortex. Interestingly, CAS has little to no effect on the amplitude of wave-V of the ABR (Galambos and Makeig, 1992), which is thought to be generated in the brainstem (Melcher and Kiang, 1996). Conversely, CAS results in significant amplitude suppression of the middle latency response (MLR), which is thought to be generated central to the brainstem by neurons in the midbrain and cortex (Kuwada et al., 2002). This finding (i.e., significant effect of CAS on the MLR but not the ABR wave-V) suggests that suppression of the MLR may be mediated by mechanisms that act on midbrain/cortical neurons. Similar mechanisms likely account for the effects of CAS on the 40-Hz EFR as the MLR and 40-Hz EFR are thought to share generators (Bohorquez and Ozdamar, 2008). It is unclear how the MOC reflex could suppress the activity of midbrain/cortical neurons without also affecting brainstem neurons given that the reflex primary targets OHCs in the auditory periphery; however, as discussed below, other MOC targets in the brainstem may play a role. Second, CAS-related suppression of TEOAE amplitudes was not significantly correlated with CAS-related suppression of the 40-Hz EFR as we report here and as was reported by Mertes and Leek (2016). This finding is inconsistent with the parsimonious hypothesis that a given reduction in cochlear amplifier gain, as measured via TEOAEs, results in a proportional decrease in neural activity as measured by the EFR. Finally, the lack of an anti-masking effect for conditions with ipsilateral noise and the CAS-related enhancement—rather than suppression—of the 98-Hz EFR measured in quiet are inconsistent with the effects of the MOC reflex on AN responses. Nevertheless, the effects of click rate were not addressed when assessing the effects of ipsilateral noise. Thus, AN adaptation or simulation of the ipsilateral MOC reflex may have reduced the putative anti-masking effect. Arguments against the role of the MOC reflex on CAS-related suppression of EFR amplitudes must be tempered by the finding that the MOC system includes collateral branches within the brainstem from cells within the cochlear nucleus (CN; e.g., Fujino and Oertel, 2001). The branches have recently been thought to compensate for peripheral inhibition according to the duration of the elicitor (Boothalingam et al., 2023). This compensation may occur from complex efferent processing that results in different effects of CAS for CAP_ENV than for EFR; however, future research is needed to test this possibility.

Ross et al. (2005) hypothesized that the 40-Hz EFR results from the superposition of exogenously and endogenously driven neural activity. The proposed endogenous component originates from oscillating neural networks within a thalamo-cortical loop that become entrained to the rate of the external stimulus. These thalamo-cortical networks may be most responsive to frequencies near 40 Hz (Llinás, 2003) and be less responsive to higher modulation frequencies due to the upper limits of phase locking for neurons in the thalamus and cortex (Joris et al., 2004). Ross et al. (2005) argued that the presentation of CAS resets the oscillations/entrainment of the endogenous component and thereby suppresses the amplitude of the 40-Hz EFR. This argument was supported by their finding that amplitude suppression was achieved by presenting a very brief (∼5 ms) burst of contralateral noise or introducing a single periodicity violation (i.e., a cycle of modulation that was longer than 1/40 Hz) within the ipsilateral probe stimulus wherein no CAS was presented. Interestingly, the recovery of the 40-Hz EFR amplitude after CAS or a periodicity violation followed the time course for which EFR amplitudes increase after the start of an AM probe (Ross et al., 2002). This finding suggests that a common mechanism (e.g., the buildup of thalamo-cortical entrainment effects) is responsible for the initial increase in EFR amplitudes after the onset of an AM probe and amplitude recovery observed after the presentation of CAS or a periodicity violation.

Amplitudes of the 98-Hz EFR were unexpectedly enhanced when CAS was presented, compared to no CAS (Fig. 3). Enhancement of EFR amplitudes from the presentation of ipsilateral noise have been reported in the literature for similar frequencies (i.e., 100–110 Hz; Prevost et al., 2013; Billings et al., 2020). It is unclear whether mechanisms of these ipsilateral noise enhancements account for enhancements of the 98-Hz EFR by CAS reported in the current study. The inferior colliculus (IC) is a likely contributor to the 98-Hz EFR (Szalda and Burkard, 2005). Many IC neurons have bandpass modulation transfer functions, where the best modulation frequencies among a population of IC neurons range between 1 and 150 Hz (Krishna and Semple, 2000). Neurons in the IC are sensitive to binaural stimuli and may show enhancements for spatially separated stimuli (Ramachandran et al., 2000). This enhancement for IC neurons may be related to the CAS-related enhancement of the 98-Hz EFR (Fig. 3) given that the AM probe and CAS were presented in opposite ears (i.e., spatially separated); however, this interpretation remains to be tested. Further, differences in neural adaptation for the 40- and 98-Hz click rates may contribute to differences in the effects of CAS for these two frequencies.

Results from the current study support the notion that the auditory system adjusts to the acoustic environment and these adjustments may result from distinct processes in the auditory periphery (e.g., MOC reflex) and central auditory nervous system (e.g., changes in thalamo-cortical entrainment effects). Real-time adjustments in auditory processing may facilitate control of peripheral and central auditory gain (Robinson and McAlpine, 2009) and the grouping of acoustic features to form auditory streams (e.g., Feng and Oxenham, 2015), both of which are important for robust auditory perception in the ubiquitously noisy settings of modern society.

V. CONCLUSIONS

Results of this study exhibited that EFR amplitudes measured in quiet and ipsilateral noise are affected by CAS in ways that are not easily explained in terms of reduction in cochlear amplifier gain by eliciting the MOC reflex. Although suppression of TEOAE amplitudes by CAS likely results from eliciting the MOC reflex, suppression (quiet condition for rates between 4.4 and 98 Hz) and enhancement (IBBN conditions; 40 and 98-Hz clicks) of CAP_ENV amplitudes were not observed. The lack of CAS-related suppression of CAP_ENV is consistent with persistence of efferent slow effects between elicitor presentations; however, this possibility remains to be tested. The effects of CAS on the EFR may be mediated by physiologic mechanisms in the brainstem and auditory cortex or influenced by central branches of the MOC system. Future research is needed to understand the influence of peripheral, brainstem, and cortical mechanisms that adjust to the local soundscape and how they may facilitate speech understanding in background noise.

AUTHOR DECLARATIONS

Conflict of Interest

The authors have no conflicts to disclose.

DATA AVAILABILITY

The data that support the findings of this study are available from the corresponding author upon reasonable request.